Choose Language Hide Translation Bar

cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Pictures from the Gallery 9: Innovative Graph Builder Views
Friday, March 8, 2024
A picture is said to be worth a thousand words, and the visuals that can be created in JMP Graph Builder can be considered fine works of art in their ability to convey compelling information to the viewer. This journal presentation features how to build popular and captivating graph views using Graph Builder. Based on the popular Pictures from the Gallery journals, the Gallery 9 presentation highlights new views and tricks available in the latest versions of JMP. It features several popular industry graph formats that you may not have known can be easily built within JMP. Views such as swimmer lanes, donuts charts, Delaunay triangulation, advanced overlay plots, and more are introduced to help breathe new life into your graphs and reports!   Welcome, everybody. The pictures from the Gallery 9. My name is Scott Wise, and I'm real excited to show you another batch of advanced Graph Building views. Before we start, I like to do something a little more inspirational here. I really think Graph Building can help you make connections. I was inspired by a really old history science TV program called Connections. James Burke created it and hosted it. It was brilliant. He always would show you a bunch of unrelated events and say, \!"What do these have to do with each other?\!" And show you how change is often random. When someone does something in one area, it might inspire it being used in a totally different way they never thought of. If you ever get a chance to see these, please check it out. In honor of James Burke, I'm going to talk to you about a Graph Building Connection, and I'm going to start it the same way he did his show. What does Locating Bats in Texas have to do with triangles? Have to do with patterned into Stained Glass, and have to do with saving Landmarks like Notre Dame Cathedral? Wow! That is a lot of stuff, right? I'll show you how this all got started. I originally was just looking at some data in where I live. I live in the central part of Texas, and one of our favorite things to do is go out and see our bat colonies fly out of their homes at dusk to be good bats and run out in the night and eat all the bad bugs that bite and pester you. We're fond of our bats, and a lot of our state parks have bat colonies in a cave or under a bridge. I was able to get that data from the Texas Department of Wildlife here. I'll turn on my little headers. I can see, I've got some locations here. Boy, it looks like I've got latitude and longitude data for the location. Not just the name. I have bat population. I even have pictures. Hopefully, you can bring pictures into a JMP data table and use those as a hover label point or a marker, which is cool. But I've got some information. The first thing I did, I said, just show me a map. We've shown maps before in our talks. It's one of the first things you learn in basics. There's background maps you can use. This one is a US state county background map. You see, I put my latitude and longitude on the edges there. I've got my marker, my points, sized by how big the bat population is. Here's 50 million bats in San Antonio. Wow! Here's one that I live pretty close to. This is the Congress Avenue Bat Bridge in Austin, Texas, and it's over the Colorado River that runs right through downtown. This is a really big tourist attraction. What bothered me is I live up in this little area just a little north of here, and there's a bridge over the highway that runs through my town. It's called Round Rock. The Neil Bridge has 3.5 million bats under it, but no one goes out to watch them unless you just happen to be in the area when they fly out. I was like, where else are there bat colonies within central Texas? There's actually a graph that can really help us see that. It's called a Contour Plot. I'll bring my control panel back up here and I'll show you. Right up here, it's this density-looking icon in your icon menu. What it does is it draws, whether you fill it or you just want to see the lines, it draws these lines around, kind of projected, expected output. In this case, my output was the bat population, and I had colored by the bat population. Now, I can see it does make sense that there's red, which is a really high bat population in San Antonio. If I go further west, there's the Frio Bat Cave, that's got 12 million. Down in that area, which is about an hour and a half, two hours south of me, there are bigger bat populations than up where I live, more in the greens. That made sense. I said, \!"I want to bring more points into my graph. Maybe I can use this contour plot, instead of doing a model to help me figure out if I'm at a certain latitude, longitude within this area, what can I expect in terms of bat population?\!" I did that, and I added these other points in. I'll now unhide, and unexclude those as well. You can see I estimated out based on their contour where they were. Now I've got more points. Now, I can see in Mason, Texas, what do I expect there to be? I expect there to be 5, 800, 000 in this site. It is just a really cool way of utilizing some points. But I got looking at this, and there's all these controls. These are your element panels down here. Sometimes you can get to them by right-clicking and going under the graph. But I generally go here for speed and completeness. It's got some stuff I can do? Well, I knew I could change the number of levels, and I knew I could control how round these things were and how jagged they were. But here's this Alpha. Watch what happens when I put this all the way down. Now, I don't know how these are created, but it sure looks like they're creating triangles, doesn't it? Building triangles, building triangles, and somehow coming up with these contours in them. That really got me curious. What does this have to do with triangles? Well, I went out to the community and I looked up contour plots in the community. Wouldn't you know it, I found a great blog. It was done back in JMP 15 when there was a lot of additional elements added to the contour plot and Graph Builder by Dan Schikore, and Dan Schikore, along with Xan Gregg, are our visualization development gurus, experts. He wrote this excellent, excellent blog. If I go down to the blog, right here, said something about using triangles to come up with how the contours are formed. Here he's got a little scripted graph where he drew those triangles with the contour in the background. He even put a link in here, and this is a Wikipedia page talking about the method they were using to create these contours. Delaunay, if I'm saying that correctly, triangulation, where you take points, and you draw triangles connecting those points, and you draw a circle where the circumference takes those points, and somehow out of that, they get these contour lines. We're not going to go into the statistics on that one. I think it's quite cool. But what we will do is maybe find a way to save the triangles. The triangles are beautiful. Reminded me like network diagrams. I said, \!"I really want to save them.\!" I couldn't find a way to do it in Graph Builder. But if you go to the old contour plot GUI, it's under Graphs, again, not under Graph Builder, but this is an older one, and this produces the contour plot. You just do that by filling in the dialog. I've gone ahead and done that, and you can see it gives me the same thing I had before. But this time under the red triangle, it has some additional saves functionality, and you can save the triangulation. What does that look like? Can I use that then to create a graph? I saved it out, and here we go. It's a stacked, ordered data table with just the coordinate positions and latitude and longitude. Now, there's everything for triangle one. I said, \!"That's cool.\!" If I go to Graph Builder, I guess I can put my latitude on the Y, my longitude on the X. I can put maybe overlay by the triangle. Let's add in some lines here and let's see if we can join these up. I know these are row order lines, so I'll click on that. Close but no cigar. Take a look at one. It looks like it's only drawing the lines for two sides of my triangle. I want to close that triangle to make it a little easier to see. I figured out that all I needed to do was modify this table a little bit. For triangle one, I figured, well, row one and two is the first line, row two and three is the second line, but it doesn't know how to connect it back up and close the shape. To do that, I took the first row's coordinates and I put them in as a fourth row under triangle. There you go. Now, I have all four of those. Now, I can go and now I can close that triangle. Not only can I close that triangle, if I go back to my Graph Builder Control Panel, I can also fill it. This is what it would look like just without filling it. But there's these fill lines down here. On these fill lines, you just go fill. This time I do fill below, and now I have a beautiful-looking piece of art. This reminded me of stained glass. Reminded me of putting patterns and mosaics and that type of thing. I started to stare at it and I started to see a pattern. Do you see a pattern in this? It's like the 3D puzzles. Maybe it's coming back at you. I'm seeing a pattern. I decided to remove just a couple of triangles, so you could better see the pattern. Let's show you what that looks like. There we go. We have a beautiful bat-shaped mosaic/stained glass art. That's just beautiful to look at. Brought to you by my best guest for where bats are located in central Texas. Isn't that great? Now, here's a connection. I was sitting there like, \!"What could this be used for?\!" Obviously, we could put this in any gallery, right? That's why it's in pictures from the gallery. But would it also have better usage. I said to myself, \!"Where else could I use this?\!" Then I was reading about the reopening of the Notre Dame Cathedral. There was an awful fire a few years ago. It burned a lot of the historic areas of the Notre Dame Cathedral in Paris, and they have rebuilt it. I remember in reading the article, they were having a lot of debate over how historically accurate they should make the reproductions because they're going to reproduce some art that got damaged or burnt, some mosaics. Notre Dame's really been built over the centuries as old as the city itself. I remember the mayor of Paris had suggested, \!"Hey, maybe up toward the top, you can put something more modern in there, like maybe a nice mosaic or plaque with all the names of the donors and the leaders of Paris.\!" Of course, I'm sure he wanted his name on there, maybe re-elect mayor Pierre, something like that. But on this one, it didn't go over very good. I think they kept things historically accurate, no matter what time period the art was introduced into the Notre Dame Cathedral. But if they're open to it, I said, \!"You know what? I can give you my bat stained glass art, and it serves two purposes, and maybe it could protect it.\!" Because here you got your bat stained art and there, if you shine a light through it, what do you get? Man, you're going to get the bat signal. Batman can now come and save Notre Dame in the city of Paris from any danger you might have. Just having a little fun with this. But hey, if you think you could convince the leader of your city's big landmark to put in, to buy our bat-inspired stained glass art that doubles as the bat signal, let me know. I'll be glad to come to your city and do that. Hopefully you're not laughing too hard at this point. But we did learn two tips in doing this, and then hopefully you got a little inspiration. One tip that I think we've learned is, hey, contour plots are really useful. They are in Graph Builder, and they are underutilized, in my opinion. Also with lines, you can do fills. This is not the last time we'll talk about filling in lines, but that's also a very good visual tool to use. Thank you for joining us. Let's clear the stage and get the true star. The true star are the pictures from the gallery and here's our version for this year, version 9. We have super plots, swim lanes, survival at-risk plots, donut plots, and two that are now available in JMP 18. I'm showing you JMP 18. That is overlay encoding and page grid. I'm going to feature these charts as time allows. Let's take the first one. Now, I want to point out that my gift to you, I'm going to give you… Like I do every version, I'm going to give you this journal. When you go look at the abstract, you'll be able to download this journal. In this journal, you will have the pictures. You will have my talking points. You will have, more importantly, all the steps to make the graph. You will have a link to which has the embedded data with the ability to reproduce it by clicking on a script. You can practice this yourself. You can get these same views yourself. Let's talk just a little bit about this view. Super plot, this is a popular chart that is starting to show up in the biotech research field. What it is, it's really an overlay chart. It allows you to put raw data. Then you could put the information about group data statistics on top of that and maybe another layer over a group statistics. You can get group means, and group, group means. It does it in a pretty easy way to view, and that's why they call it the super plot, a lot of information on one chart. How can we create this? I'm going to follow my steps. Here, you can see my data set-up. I have two treatments, I can control, maybe where the real treatment happened. I have over three groups. I have the real data. Then we have a formula column where all I did was ask for the column mean of data by treatment by group. All that is ready for us to use. I'll click on the Graph Builder, and let's build it together. I'm going to take both sets of data together in the Y-axis, I'm going to put the group down into a treatment first. Let's put treatment onto the X, and let's put group in the overlay. A little confusing. I'm not so clean right now, but that's okay. We'll clean it up. I do need to right-click in here and add a second points element. Now I have points on top of points. Little confusing until I start to make some selections. What I'm going to do right now, this first points, I want this to be the raw data. For this first points, all I have to do is open up this variable selection and remove the means. Now this first set of points is just showing my raw data dispersion across treatment, which is cool. I will close that one. The next one up here, this is one I want to do a little different. Instead of doing it by the data, I want to do it by the means group. Now, I've got many points all showing the mean. I can put that jitter label all the way down in this area. I can also change the summary statistic just to be the mean. Now, I've got it here. Now, it's a little hard for me to see. I can right-click right into the graph, I'm going to do this. I'm just going to go to Graph, and I'm going to do Marker Size. If you want to change the marker size of all the points on the graph, you can do it this way and make it really big. Now, what can I do to clean up this view? This is a good time to go under the red hotspot, the red triangle, I call it the hotspot, and right on the Graph Builder menu and pull up your Legend settings. Let's add back that second set of points. Now, let's work with not just colors, which we can't change the colors, but let's work with the markers. Maybe I can change the markers to make this cleaner. Here under A, I can go under Marker. Instead of a little solid dot, let's make it a triangle. Under the other A, I'll make that one an open triangle. That's the mean of the groups, and that one's going to be the raw data. For B, let's not do points again. Let's do a marker of a closed triangle. For the B on the second set of points, let's do an open triangle, and so on, and so forth. We'll do diamonds here. We'll do a solid diamond for C, for the first point, and for the second point, we will do a closed diamond. There we go. Marker Size, closed diamond. There we go. That's looking a little better. I know on the first set of points, you can also spread out how far those are looking, and I might spread them out just a little bit more. I think one thing that could help me, I could right-click here into the X-axis, go to Axis settings, and I can add a darker line here right under Control and Treatment. You can see where the group mean of A is under treatment and control. That's that hollow circle. You can see where it is under B, and then you can see where the rest of the points fall. That's nice alone. But can we beat this? Yeah, we can. We can actually bring in another level of information. I'm going to add Bars, and this is going to add a bar chart. Now, that's not a very helpful bar. I would love to see the mean of the treatments irrespective of group. To do that one, I'm going to change the style to a float. I might turn on some error intervals, look at my standard error around that. Now, you see it's got broke out by group. Obviously, in variables, I need to do some things here. I can take out the data, I can take out the overlay group, and now I've got a pretty nice-looking chart. I'm going to say Done. Now I've got really three levels of information on the same chart. Pretty easy to see. As well, and I'll show you this probably once because I do this a lot, you can bring in pictures as backgrounds, and I do this every time I do pictures from the gallery. I've got this image as a background. Now it's an image you can right-click, you can say, Image. Let's size and scale with the fill the graph. Oh, that's drowning out my data, so let's make it transparent. Maybe make it 0.3 or something. Now, that's a nice background picture. Of course, you can change your titles, you can change what your legend settings mean, the position of the legend settings. But that's the whole idea. You have that available to you. That is the super plot. All right. Very cool. Well, let's move on to the next chart. The next one is Swim Lanes, and I owe a big thanks to my good friend, Jed Campbell. Jed is very talented. He's extremely talented statistician and JMP guru and custom scripter, but he's also a good woodworker. He's got all these talents, and he was making this beautiful... That is a bowl that Jed made. He featured this in a blog, and I put the link to that blog, so you can visit. He actually created a wood turning segment calculator in JMP to help him make it, and he created this view with swim lanes. Swim lanes are like what you think. If you're in a swimming pool, you've got your own lane to swim in. You can go and swim down that lane and no one else is going to swim in your way. This is creating those type of swim lanes in your graph. This really helped them for a ring position as it gave them these inner and outer radius dimensions. What we can do here is go pull up this data. Ring is a ordinal data set. It could have been categorical, but you need something ordinal or nominal to make that work. Then he's got these inner and outer dimensions. Those are continuous. Let's go ahead into the Graph Builder. Let's put that inner and outer diameter on the X-axis. Let's put Ring into the Y-axis. Okay, so far, so good. He got points. I think he would also like to have lines. And not just lines, he wanted to have a step line. The connection type here will change from line to the centered step. Oh, that's nice. It looks like that's doing well within the swim lane, so you can fill it. This would be a fill between this time. It looks like the graphics doing what I wanted to do. I'll just make this stuff a little darker on our screen. But take a look at when I start to add grid lines. It's not the fault of the X-axis. That one looks pretty good. But if I go right now under the Y-axis and I go to my axes settings and I try to add the grid line, look what it's going to do. You can see it in the preview alone. It is going to put a line right by the number of the ring. That's not what I want. That's not the swim lanes. The ropes are not where I want them to be in my pool. To do that, if I just go back to Axis settings, Jed showed me, you know what? If you click right here where it says Show Tick Labels, instead of just doing marks, you can do dividers. Here's short divider. Take a look at that. That makes the swim lanes. You say, Okay, and now we have the right view. Now, you can bring in the picture of his beautiful bowl, but this was actually what he needed to help him make that bowl and have the right swim lanes to make it easy to view. All right, so Swim Lanes. Thank you, Jed, for that contribution. Please do see his blog. Overlay Encoding. It's really funny. I was learning of new things you could do in JMP 18. There were times before JMP 18, like in 17 and the previous versions, that you would do too much stuff. You would add too many elements on your graph. In the overlay element, if you use that element, it would take over the colors or the style of what's being presented. They fixed that in 18, and I found the perfect data to help us see that. It's going to show off what we can do with overlay encoding. This is ice accumulation on the Great Lakes between Canada and the US. This is one of the things they're looking at that's helping them measure the impact of climate change in the warming planet. Is there less ice over the winter periods? This is like Lake Michigan, and you can see I've got to broke it out in two distinct time periods, a before and after. Well, let's go ahead and pull this data up and make it work. Here's this Great Lakes Ice Accumulation. We'll just do it for all the lakes for this example. We'll go do our Graph Builder. There is our Graph Builder. Looks like I want my ice accumulation on my Y. I want a month. That's ordinal, and it's ordered ordinal. I've I'm not sure that 11 started first because these are the winter months, at least up near the Great Lakes. I don't consider June to be a winter month, but I guess at some point they had ice still in the Lakes. All right, that's cool. What else do we want? I want year. We want to have separate points and lines by year. That's an overlay. I have overlaid by year. Now, I can right-click in here. I don't want box plots. Change me to box plots. I want points, and I want to add the smoother lines. Now, under points, I'll go to these graph elements. Now, let's go ahead and make those means, which is cool. Down here, I might turn on this Confidence of Fit for the smoothers. All that looks pretty good. If I take a look at it now, it's busy on the legend. I might do away with the legend, but you can see my points and my lines are the same color. Well, what if I wanted to bring I've got another element? I've got this winter period, and it's going to show me the difference between times after 2016 winter season and times before that winter season. I like to bring in that data. Oh, well, that's no problem. Just take winter period here and throw it in the color now. Oh, no! What happened? It took over the year data points, and it started to give them a symbol. Did I really want it to do that? I don't know. But now in JMP 18, you have a place. It's right in the overlay landing drop zone. You right-click and you say overlay encoding. It's got an auto, and I think the auto is always... It decides between color and style. It's usually color. But let's see what color would be. I don't like color because now my lines are the breakout among the two different distinct periods of years. But now it's given different colors that I can barely see. I mean, I can go under Graph and try to make these marker sizes bigger, but I have a hard time seeing the colors, and I want everything to be blue or red. Well, how do I do that? I'll right-click in overlay encoding. Maybe I do it by style. No, that is silly. Right click. Oh, my goodness, there's a none. I can turn off that control among the overlay and just let it overlay by it, but let the other landing zones, in this case, the color landing zone, come forward and take over what color things should be. This one's proper. I don't need to keep all this other information in the legend. I might drop that, but that one's given me the proper view. So that easy to do. If you take a look at the one I've encoded in here, I put some axis lines. I have put a little filter here, and now you can play around with different… Lake Michigan, Lake Ontario, a place I used to live, Lake Erie. See if you're seeing the same trend within each individual lake, and you can sail your little icebreaker through the ice. All right. It is showing, I hate to say, but it is showing that we are entering a period over these last eight years of less ice accumulation, which means the temperature is warming. All right, very cool view. We thank development for that overlay encoding. We can do now more types of graphs. Okay, about halfway through, we're doing pretty good. I think we're good to see the next set. In the next set here, this is a challenge one. Just like the super plot, our customers came by and said, \!"Hey, I want this view.\!" This is a very typical view for when people are doing clinical trials, and it is getting a survival plot with the at-risk numbers below. I have this graph showing me survival among these two groups. I even have a little aligned indicator of how many of my populations' left. One person in the test survived out to this time period, which was a thousand. One person didn't even make 600. I'm on the blue one in standard when there was probably between four and three people left. This is a really nice view. This is actually stacking a bunch of graphs in the same window that makes this happen. It can be created. Now, if you have a product like JMP Clinical, this type of report is the type of thing that gets built into those automatic clinical reports that help you explore and analyze and basically report out your data. But you can do this in JMP, and I'm going to show you how. Where most people would go first, if you don't have the data set up yet, if you just We have basic raw data like we have here, there is a nice platform, might not be where you think it is, but you can make a survival curve. For those that are interested, it's Analyze, Reliability, and Survival. Here it is. It's called Survival. Time is my time. Treatment is my grouping. When you do these kinds of charts, you might have an indicator over whether something at that time frame survived or didn't survive. That's called censoring. You see that in reliability, too. We'll click on there, and this is just to make that graph. That should look pretty similar. It's a jaggard graph. It's not smooth, but it's the same thing. But where are the at-risk numbers? Well, they happen to be in these bottom three tables down here. If I open one of them, you can see there's the at-risk number. But it dawned on me, Oh, I still have time and survival information in that table. One thing you can do anytime you have reports open and there's tables, you can right-click and make a combined data table, and it will stack any like data tables that are open in that report window. There's the standard, there's the test, there's the combined. I'm really interested in standard and test, so I saved that out. That's where I got this data, and that's where you can get the information to make your survival curve. If you don't like the one that's already there, we can go ahead and make that. I've already saved that. When you look at your journal, it's going to be the VA Lung Cancer Combined. This is the one, and I just kept the standard and the test from the stack. Now it should be relatively easy because now I have my graph data in with my table, my at-risk data that I can make into one graph. Go to Graph Builder. Now I'm going to put Survival on the Y, Time on the X, and let's use the grouping for the overlay. Let's remove the points that I have smoothers. In three clicks, I've regenerated that view. That's pretty easy. Now, below it, maybe I can take the group. Now, remember, the group was like my treatment, whether it was standard or treatment in terms of my treatment types. I don't put it in with survival data on the Y-axis. I put it below. It makes a split chart, which is cool. You could keep these separated. There's also a little red hotspot setting here which does graph spacing. Now that I've got two in here, I can make the line in between them a little thicker. That's just a little fun thing. But I'm looking at this now, it's got points. Well, that's not the numbers. How do I do the numbers? To do the numbers, I will go under this little red triangle options under the points element. I'll say Set Shape Column. I'll make sure the at-risk is selected. Now I have the at-risk numbers. In fact, I'm going to right-click in here under a Marker Size and make them big. That's cool. You can jitter points, and it jitters them up and down, and I can put all the way to the left, and now they're all on one line. Now, this is more data overlaying on top of each other to be really clear. I might want the filter on that. But that's the view that I'm looking for. You might say, \!"Oh, this looks terrible at the bottom. How can I make it look a little better?\!" If I take my pointer into the axis, it becomes this little grabber hands, and I can move stuff up, and I can move stuff down in this area, and I can just work and even get the little grid line up here, and I can just work and massage until it becomes a little smaller than the little smaller space and then the one up top. I like that. Now, all I have to do, if I say Done, is go to my local data filter right there under the big red triangle menu, red hotspot. I'll just put time up here, and now I can just do select times. I can start at zero. I can put some at eight. I could put some at 18. I try to do them every 10 or so and start to build out this graph. I've already done this. It's actually scripted in here into the data. I brought in a picture and add some axis lines. But here now you can get just the time periods you want lined up with that same scale on the graph. When you add a axis line to your graph, and I've added these grid lines, it works for both graphs because they're really the same graph. You just stacked them on top of each other, sharing the same. They're two different graphs sharing the same axis in the same window. That's a pretty cool view, and I'm glad we were able to meet that challenge. Maybe that will be helpful for you as well. In anything you're looking at, something performing over time, this might be a great chart to look at. It doesn't have to be clinical or biotech data. It could be warranty failures. It could be anything out there in the field or out there in a lab. All right. That was our survival at-risk plot. Let's take a look at page grids. Here is something else that we can only do in JMP 18. What it is, is sometimes when you're comparing a bunch of charts in a grid, you would like them to maintain their individual axis scales because you're just looking for trends. You're not trying to combine. You're not trying to compare the level, the magnitude of the numbers. You're just looking for trends. This was really hard to do before without some probably customization. I'll show you what this looks like. I have all these coffee export in countries. It has how much they export domestically or consume it as well, has all this information. This was a really good one to show this. I will open this up. I've got my Graph Builder. I'll take my three outputs and put them all on the Y. Instead of points, I will change to a line. I will put it down by year. There's my years, looking over growing seasons, obviously. I like this, but I want to see it by the countries. I've got this Page button down here. Well, let me create different little views by page. You see it just stacks them on top of each other. I'm going to say Done here. I'm going to right-click, and I am going to go and turn on a local data filter and let's just look at a two-by-two grid of some of my favorite countries. Brazil, great coffee producer and consumer. Costa Rica, get a lot of coffee from Costa Rica. I like dark roast. They have good dark roast in Costa Rica. I got to put my second home, which is the Philippines on there, and let's do Vietnam, one of the up-and-coming coffee producers. Now, I have these views. You can right-click up here and say Levels Per Row. Let's just do two. Now I have a grid. This is nice because every country has their own individual axes. Now, before JMP 18, this is what you would have seen. There we go. This is what you would have seen. It would have forced everybody to share the largest scaled axis, which is Brazil, which goes from zero to 50,000 or above. Well, Costa Rica, you can't even see a trend. Philippines, you can see a trend. Vietnam, you can see a small trend. But for trend examining, it doesn't really help you. You don't have to come in here and force something as well. If you right-click right in here, you can go to this Link Page Axes. This is new in JMP 18, and you can say, I want, if you have this Replicated Page Axes clicked, it will go across your grid, and it will allow you to say, hey, share X-axis, share Y-axis, or share none. In this case, X-axis is the only thing that these guys share because everybody had the same yearly periods. Now I've got it, and now it brings it back to where the default was smart enough to see. Now I can see trends irrespective of the numbers. I know Brazil is 3X, 5X, 10X, the others. But I can see your downward trend for Costa Rica now, flat consumption, a little increase in consumption. Oh, my God! Look at the Philippines. Look at that. Starting in about the last part of this data set, their domestic consumption, which is the green line, has gone up. Now, I'm going to visit there pretty soon, so I'm going to be in good shape because if there's something I like, I like my coffee here. I'm going to have fun drinking the native Filipino coffee. You can see the Vietnamese trend on how they're really going up at a steeper slope, in this case, in terms of exporting and total production. This is just a really nice way to control a grid view. Of course, the one I put in here has pictures, and I put a nice picture in from each country, so you can make it even more beautiful than it already is. All right, we are almost at the end. I think we're doing pretty good on time. We will go look at the last plot. It is Doughnut Plots. This is a version of the pie chart called a Ring Chart. I think it's a useful version of the pie-type charts. If you don't have too many categories, and they all have enough of a presence, it makes a great way to select filter for another chart. I got this really cool data to help me show this. I'm going to teach you how to make a very easy dashboard. If we click on something like glazed doughnuts here, I want it to change my view for how the breakout is among the generations. Can we do that? Like Homer Simpson, am I making you hungry with doughnuts? I probably am, so I apologize for that, but hey, who doesn't like doughnuts, right? I got this great generational doughnut data. They did a big study during the pandemic. Don't know how they got everybody, I guess, online. They went and asked them what kind of doughnuts do you like? Then they got all these percentages out of it. Baby boomers all the way up to Gen Z. I keep mentioning in here, I put a value order here. I could custom order and baby boomers will show up at the bottom of my charts and that type of thing. I did put an order to this. Now, makes it really easy. We're actually going to make two charts. First one is going to be that, aforementioned Pie chart. I'm going to put my Type. Just Type on the X axis, I'll put my Percent on the Y. Here, I can right-click and change out to a pie, or you can just select the pie symbol up here. I don't like that type of pie chart. I'll do a ring chart. Since I have this percentage data, I can go ahead under labels and label by percentage of total values. That's simple. Overall, 30.7% like glazed doughnuts. That's what a glazed doughnut looks like. I've made the labels look a little bit helpful. Chart 1, there's my ring chart. Now, can we do something with the bar charts? I'll go to Graph Builder. We will do a bar chart here. To do this bar chart, looks like I'm going to put a Gen on the Y Percent on the X. There's Gen on the Y, there's Percent on the X. I'm going to ask for bars. It doesn't look too interesting. I'll ask for the sum right now. Under this percentage, I'll go to the Axis settings, and you can give it Percent style. That looks really cool. Boring chart right now, but maybe this would be a cool chart to link to it. If this could control the settings in this one, that would be awesome. Well, I knew you could go under File, New Dashboard, and I knew there were Templates and you could create this from a template, but there's an easy way to do this. Go under Windows, go to Combine Windows. Most people don't know about this. It's showing you here some data. We don't care about the data. But here's that first graph. Yes, I want to combine that one with the second graph. On the first graph, hey, filter by that graph. You say, Okay. Now you have a ready-made dashboard. In fact, we can get rid of the others just to show you it's its own thing now. If I did this right, I click on the glazed doughnuts. Oh, my goodness! It changed how it's looked at by the generations. So cake doughnuts. I love cake doughnuts, but I'm a Gen X, but even baby boomers ate more. But cake doughnuts are really good for dunking, right? Remember Dunkin' Donuts? That's where they got the name. You dunked your doughnut in your coffee. It was really good. It made for a soggy doughnut. That's why you needed a good firm doughnut like a cake doughnut. You wouldn't dare do that with a glazed doughnut. But it doesn't seem to be as generationally popular now. Let's see, everybody likes these cream filled, and these jelly filled doughnuts. What about sprinkles? Oh, Gen Z love sprinkles. I like sprinkles too, but that's just a kid in me. Everybody likes sprinkles. Other, I mean, that's the weird voodoo doughnut, right? Putting the bacon on top. But I won't buy glazed doughnuts. The most famous glazed doughnut is the Round Rock doughnut. That's the only tourism we're known for. You can come to our Round Rock doughnut bakery and get a huge-sized glazed doughnut. Texas size. There's even one up here where I don't eat. You can see I gave it the skull and crossbones because you can't spell diet without die. It's just a saying. You must be dead if you're not enjoying doughnuts. But unfortunately, people like me and Gen X and baby boomers, we have to worry about our weight a little bit more. So maybe we're not eating doughnuts now. Terrible. I still am. That's cool. You could save this as a script. This is the easy way to get into a dashboard and use a ring selector. Hopefully, you enjoyed that. All right, so that is our pictures from the gallery. I will leave you behind as well with places to see our other galleries out on the community. There's a master list of all the ones we've done starting in 2016 with links to where you can go get the journals for those, so you can replicate them. Go see waterfall charts, all kinds of great stuff. There are things here, and this one will be added to that list as well. We have other blogs and journals we recommend. I also have the other link to JMP's blog up in there, lots of good graphing ones. There are good as well presentations that Xan Gregg, Dan Schikore, Dan Valente, really smart visualization people have put in here. There are other tutorials available on the community as well as just go to the learn jump in the community, so you can see all the training that we offer in terms of visualization and graph building. If you have a challenge, which you can't figure out how to do in Graph Builder now, or you want to know if we would add a different element, a graphing element, please go to our JMP wishlist on the community. You can search to see if it's been asked before, and you can give your own wishlist. This is a way for us to get the voice of the customer when it comes to developing new views on graphs. All right, so I'll leave you behind with our beautiful pictures from the gallery. I thank you for joining me today, and we hope everybody has a great discovery. Please go out and play, explore, connect, get inspired in Graph Builder. Thank you.
Event has ended
JMP Live is for You, Too: How to Use the Content that Your Colleagues Publish to JMP Live
Friday, March 8, 2024
In this talk, we cover tips and strategies for making use of the data and analyses that others in your organization have published to JMP Live. Topics include: Finding content that is meaningful to you (browsing, searching, and sorting). Finding the data used by a report, finding the reports that use a given data set. Bookmarking content once its been found. Interacting with reports (using explicit interactive features, hovering, excluding certain data). Discussing reports and data in comment threads. Calling your colleagues' attention to the content (sending links, @ mentioning). Downloading content to explore further in JMP.   Hello, and welcome to the JMP Discovery Talk titled JMP Live is for you too! My name is Aurora Tiffany Davis. I'm a software developer on the JMP Live product, and I created this talk to show you how JMP Live can be useful to everybody at your organization, not just those people who produce the data and reports. In fact, in this talk, you'll see that for most of the things on JMP Live, you don't even need to have JMP on your machine. Whether or not you have JMP installed on your machine, you can browse, filter, search, and sort the JMP reports that your colleagues have published to your organization's secure JMP Live site. When you find something interesting, you can bookmark it. When you open a JMP report that's been published to JMP Live, you can not only look at it, but in most cases, you can also interact with it in many of the same ways that you can do in JMP itself. You can read and join in on the conversations that are taking place about the analysis at your organization, and you can @mention a colleague to draw their attention to something. You can use JMP Live to learn more about JSL scripting. You can use JMP Live to learn more about the relationships between JMP reports and the data that they rely on at your organization. You can download content from JMP Live. Okay, let's get started with a tour of JMP Live. Now, I'm not going to show you everything in JMP Live, but I want to go over the basics because I want to make sure that you know how to find your way around, how to find out what you can do, and what you can learn. We'll start our tour here at the top with this blue navigation bar. This bar is always present no matter what you're doing and where you move around within JMP Live. We can see that right now we're on the homepage. On the homepage, you see a list of all of the reports that you have access to. In other words, all the reports that are published to a collaboration space that you are a member of. You don't have to worry that everybody in your organization can see every report. That's governed by which collaboration space the report is in. The homepage is a great way to see at a glance what's new because the default sort order is newest first, so new content is going to bubble up to the top here. Now, any place on JMP Live, like this one, where you see a list of reports, you can click on a report to open it, or you can bookmark it to come back to it later. Now, I'm not going to open this report yet. We're going to get to that a little bit later on in our tour. For now, let's look up at the top of the homepage, and I see this top bar. When you see a top bar in JMP Live, it's usually telling you two things: where am I, in this case, on the homepage, and what can I do? What can I do on the homepage? I can filter my results. Maybe I don't want to see all of the reports that I have access to. I can filter according to the text that's in the report title or description. I can filter by which user published the report. I can filter by collaboration space. I can filter by date range, or I can choose just to look at control charts that have active alarms right now. What else can I do on the homepage? I can change my sort order. I mentioned a moment ago that the default sort order is newest first, but maybe instead you want oldest first, or maybe you want to sort alphabetically. What else can I do on the homepage? I can change my display format. Right now, the display format is Grid, which is the biggest format. It shows you a really good preview of what the report looks like, but you might want something more compact instead. Okay, that's a quick overview of the homepage. Now let's move along on the blue navigation bar from Home to Spaces. This shows you a list of all the collaboration spaces that you are a member of. Your organization may have many, many spaces, some of which are really of no interest to you at all, and so we don't show you those spaces here. If I look at the top, I can see where am I, and what can I do? I can search for a particular space, or I can filter the list of spaces that I'm seeing here. I can also click on any one of these collaboration spaces to open it up and take a look inside. Let's open up the Reliability space. When you open up a space in JMP Live, you see a folder tree showing you all the folders that are inside that space. When you click on a folder, you can see all the reports that are in that specific folder. Like any other place where you see a list of reports, you can click on a report to open it, or you can bookmark it to come back to it later on. If we look up here, we see a top bar. Anytime I see a top bar, I know that it can tell me where am I and what can I do? Some of There are many things I can do are familiar to me from the homepage, for example, changing display format, filtering, and sorting, but there are a few things here that are new. For example, I can email a link to a colleague of mine, and this will be a link to this specific folder. I can download a JMP Live folder as a JMP project. This will download to my machine a self-contained JMP project that has a copy of everything in the JMP Live folder, so it'll have a copy of all the reports as well as the data that they rely on. In addition to this top bar, I'm noticing that over here on the left, we also have a sidebar. Whenever you see a sidebar in JMP Live, you should think, what can I learn? What can I learn about a space? I can learn about the posts, the report posts, and the data posts that are inside. That's what we've been doing so far on this space. I can also learn about control chart warnings in this space. Moving along from Spaces, we'll go to the Sitewide Search. Maybe there's a particular project going on in your organization that you're interested in, but you can't recall the name of the person who's been publishing reports. Maybe you don't remember what collaboration space it's in, all you know is the name of the project. This is where you would go to find more. If I'm interested in the Twinkle Tech project, I can type in Twinkle and see a list of reports, which again, I can click on any report in the list or bookmark something to check it out later on. I can also search site-wide for users. Maybe I'm interested to see what Dieter has been up to lately. I can find Dieter easily, and I can click on his name to open up Dieter's user profile. The user profile shows you everything that they have contributed to your organization conversations JMP Live site, or at least everything that you're permitted to see. Up at the top, I can see, again, where am I and what can I do? I can filter, sort, and change display format. I have a sidebar, and I know that means there's what can I learn? I can learn about reports that Dieter has published and also the data that Dieter has published. Moving along from the site-wide search, we come to Notifications, and I can see that I do have one notification. It looks like Michael Goff has mentioned me in a comment on a particular post. I could click on this to open that post up and see what Michael said. This is also where I would change my notification settings. For example, whenever I'm mentioned in a comment, I certainly want to be notified of that. In fact, I also want to get an email about it immediately. But there are other notification types that I'm less interested in, maybe I don't want an email. This is where you go to set up those preferences. Moving along from notifications, the next thing on our blue navigation bar is Bookmarks. During this tour, we've been bookmarking things here and there, and this is where we can go to see a list of all the things we have bookmarked. Up at the top, Where am I? Bookmarks, and what can I do? Filter, sort, and change display format. Moving along from bookmarks, we come next to Help. This is where we have a link to the comprehensive online documentation for JMP Live. During this talk, I'm only showing you some of the things of JMP Live. If you go to this online help, you really see everything. Next, we come to this last section on the navigation bar. This is links and settings that are related specifically to you. This is where you can go to easily open up your own user profile. This looks pretty much just like the user profile that we saw for Dieter. It shows me which reports and which data I have published to JMP Live. Also, in this area is a link to your own personal collaboration space. This is just like any other space, except it's for you specifically. You can publish much content from JMP into this space, and you get to decide who else is allowed to see it. Also, in this area, we have various other settings you can apply. Maybe you want to look at JMP Live in dark mode, you're going to change your display language, that thing. This is also where you can sign out. Okay, now that we've finished our tour of the navigation bar, let's get to the really good stuff. The whole point of JMP Live is sharing interesting sharing JMP content with each other and collaborating. Once you find something that you think is interesting, click on it to open it up. Here, we're opening a report. This is a JMP report that somebody has published to my organization's JMP Live site in this particular report is a map showing the last six months' worth of building permits that were issued by the city of Raleigh, near JMP's headquarters in the US. In most cases, when you open a report in JMP Live, you're not limited to just looking at it. You can also interact with it. When I first opened this report, it's showing me all of the building permits for the last six months, and it's a little busy and a little hard to digest, but the publisher has included an explicit interactive feature in the form of a local data filter, so I can very easily change to just looking at the residential permits. You can also interact with a legend on a report. This legend is showing me various types of building permits. For example, new single family dwellings, also known as new houses are shown in pink, according to the legend. But there's so many points still on my map that it's a little bit hard for me to tell where new houses are being built. If you click on a category within a legend, it will highlight those points in your report. Now I can highlight just the new houses and really easily see where they're being built. You can also hover over a point and a report to learn a little bit more. Now, all this interaction I'm doing: the local data filter, clicking on a column in a legend, hovering, this is my way of asking JMP Live to show me a version of the report that I'm interested in. This is not going to affect what anybody else is going to see if they come to the same report. Also, I'll note that you don't need JMP on your machine to do any of this. All the work that's taking place to facilitate this interaction is taking place on your organization's secure JMP Live server. Let's look up at the top here. I see we have a top bar on a report, which means I can learn where am I. This is a path to my current location. I see that I'm looking at a Map Report and that it is inside of a folder called JMP Live is for you too. That folder is inside of another folder called Discovery Europe 2024. All the pieces of this path are clickable, so you can use this to navigate around and get a sense of where you are. Another part of the answer to the question, "Where am I?" Is in this area here. This is an area where you can navigate back and forth among the reports that are in a folder together. I can see that this map Report is not the only thing in the folder. There's a previous report as well. Let's click this and take a look at the other report in the same folder. This one shows how long it took for the city of Raleigh to issue building permits of various types. Just like in the Map Report, I can click on a category in the legend to highlight points in the report. Just like in the Map Report, I can hover over something in the report to learn more. Unlike in the report, I don't see an explicit interactive feature included by the publisher, so I don't see a local data filter or a column switcher. Nonetheless, there is more interaction that you can do with a report on JMP Live. I'm not very interested in the building permits that were issued quickly because I think that should be normal. I'm going to select the building permits that were issued quickly, and then I will right-click and say Exclude and Hide Selected Rows. Now I'm seeing just the data that I'm interested in, the building permits that took a long time to issue. We might want to research these and find out what's taking so long. We already talked on a report about the where am I section. Let's talk about what can I do. I can bookmark. I can share link. You can even download just one report as a JMP project. We talked about doing that earlier with an entire folder. You can also do it with just one report. I also see a sidebar over here, and I know that a sidebar means what can I learn? What can I learn about a report on JMP Live? I can learn various details such as who published it, when it was published, and more importantly, when it was last updated. I can also learn about the conversation that's going on about this analysis in my organization. I can join in on that conversation by adding a comment, replying to someone, or @mentioning a colleague so that they'll get a notification, just like the notification that I got when Michael Goff mentioned me. What else can I learn about a report on JMP Live? I can learn about the script that generates it. This is actually a fantastic learning opportunity. If you see a report on your organization's JMP Live site that's doing something really interesting, maybe somebody's using a technique that you're not familiar with, or you just find the report really attractive, you can learn how they built that by coming here to the script area, you can copy the script or download it to your machine to dig in and learn more. What else can I learn about a report on JMP Live? I can learn what data it relies on. You can also click this link, so you can move back and forth between the data and the reports that it services to get a better idea of those relationships. Let's do that now and open up the data post that that report relies on. You're seeing here a preview or sneak peek of the data. Up at the top, you see "Where am I?" It looks like the data post name is a little long, so we didn't have a chance to show you the entire path. There just wasn't room. You can click on this triple dot to see the rest of the path. Again, all of these are clickable, so you can use these to move around if you want. We also have this what can I do area, including downloading the data table. We have a sidebar, so we know, what can I learn? I can learn details like who published it, when they published it, more importantly, when it was last updated. We have a comment section, just like we did for a report. Here's something that's new that isn't present in a report but is present for a data post. We can learn about the columns that are in the data table. For any column, we can click this Info icon and find out more information about the data type, modeling type, and so on. We can also choose to just look at some of the columns. Let's deselect some of these columns using this checkbox area and then click Apply. Now we can get a more focused look at the data before we possibly decide to download the whole thing. We can also learn for a given data post which reports rely on it. We can see that this data post is relied on by two different reports. Again, these are links, so you can use this to move back and forth between reports and data. Anytime you land on something in JMP Live that has a sidebar like this, I really encourage you to take a moment to explore the sidebar because there really is a lot that you can learn about what's going on in your organization. Okay, at this point, I think that we've hit all the highlights. We've talked about finding your way around using the navigation bar and the various other links that are in JMP Live. We've talked about interacting with reports and finding out, where am I? What can I do? And what can I learn? I hope that you can see that while JMP Live is super useful to the people that produce the data and reports. It's not just useful for them. It's not even just useful for people who have JMP on their machine. It really can be useful to anybody in your organization. We think that the easier it is to share knowledge, find knowledge, and collaborate, the more effective you'll be in driving your projects forward. JMP Live can help you be more effective. I want to thank you for joining me today for this talk, and I hope that you'll also check out some of the other JMP Discovery talks that we have this week. Thank you and have a great day. Bye-bye.
Event has ended
告别试错时代，DOE 工作流程助力 AMAT 创造价值
Thursday, May 30, 2024
Conference Room 1,2, or 3
Speakers : 杨嘉易, 工艺支持工程师，应用材料 The speech topic : 告别试错时代，DOE工作流程助力Applied Materials 创造价值 Speech abstract : 软件，如利刃，在匠人之手，可雕琢万物，亦可改变世界。 来自世界领先的半导体设备制造公司AMAT，携手多个 6Sigma 真实案例，为您展现软件的力量。 案例源自实践，价值创造，值得借鉴。 AMAT 联手JMP 团队，精心甄选多个真实案例，涵盖 JMP 软件的 DOE、多元分析、质量和过程分析、消费者研究等强大功能，不仅适用于半导体行业，更可为其它行业，例如：制造业， 医疗保健，金融和服务行业提供借鉴。 打破思维定式，激发创新灵感。 通过学习和应用这些案例，用户可以掌握 JMP 的使用方法，解决实际问题，并创造价值，更能打破传统的思维定式，激发创新灵感，为不同的的组织创造价值。 案例呈现清晰，脉络分明，助力实践。这个演讲通过巧妙的工作流概念将不同的案例连接起来，不仅有利于理解案例内容，更重要的是可以帮助用户学习如何将 JMP 应用到实际工作中。 JMP，数据驱动决策的最佳利器。 作为功能强大的数据分析软件，JMP 可以帮助您做出更好的决策。这个演讲从设计实验出发，推动了整个JMP平台在推动“数据驱动决策”方面的应用，为6Sigma项目提供了关键的支持    
Labels (1)
Labels:
Event has ended
JMP DOE 分析及建模优化在预测外延硅生长速率中的应用 - 应用材料
Thursday, May 30, 2024
Conference Room 1, 2, or 3
Speakers : 沈佳苹, 工艺支持工程师，应用材料 The speech topic : JMP DOE 分析及建模优化在预测外延硅生长速率中的应用 Speech abstract : 在半导体制造中，外延硅被广泛应用于晶圆衬底、pMOS SiGe、nMOS SiP、沟槽填充等应用。外延硅层的生长速率会受到多个步骤、多个工艺参数的影响。为了更高效地建立此种强交互的预测模型，我们需要彻底且全面的DOE评估，设计高度正交的DOE。此项目以外延硅生长速率的历史数据为起点，进行设计评估。对于建立预测模型而言，该历史数据结构功效弱，D 效率低，设计均匀性差，效应相关性和预测方差较高。在建模期间，由于缺乏模型自由度，RSM的逐步算法不稳定，导致模型重复性不佳，最优设计点的置信区间过宽。除此之外，RSM 模型中观察到了两对强相互作用：一对相互作用可能归因于各效应之间的高度相关；而另一对相互作用则与工艺的竞争机制相关。随后利用稳健设计和蒙特卡罗模拟进行公差设计，利用设计空间刻画器用于进行公差分配分析，以模拟未来的技术需求。为了以最小成本改善现有的DOE结构从而优化预测模型，我们未采用全新的DOE，而是利用增强设计方法，通过三种方法改善现有DOE结构：(1) 移除非正交的数据； (2) 默认增强算法； (3) 中心增强算法。通过非常全面的增强设计，JMP建议了最佳四个增强数据点，最终改进了DOE 结构，极大地提高了对于历史数据的利用效率，显著缩短了工艺开发的周期及成本。  
Event has ended
JMP 脚本编辑在生化实验室的应用案例分享 - 诺和新元
Thursday, May 30, 2024
Conference Room 1,2, or 3
Speakers : 屈颖奇, 高级研究员，诺和新元（中国）生物技术有限公司 The speech topic : JMP 脚本编辑在生化实验室的应用案例分享 Speech abstract : 在生化实验室中，使用 JMP 脚本语言可以为研究员个人、实验室、跨部门三个层次的工作助力。研究员如果有大量的同类型的数据需要处理，可以编辑简单的脚本工具处理重复性工作,包括数据整理，做图，计算。这样的脚本做起来非常简单，每次在脚本中改变几个参数就可以用在不同的数据上，可以节省大量个人工作时间。第二个层次是用脚本工具创建实验室工具箱Addin，方便有同样数据处理需求的同事使用。实现快速数据可视化和数据分析自动化，协助研究员处理数据，调查实验中遇到的问题，实验失败的原因等，节省工作时间和实验室资源，加快解决问题的速度。前两种应用的脚本都以数据为主。第三个层次是使用JMP脚本语言创建日志模版，实现自动化撰写实验方案、数据分析，一键生成实验报告。研究员需要撰写大量的实验方案，分析数据，并撰写实验报告。不同实验室的实验员需要阅读理解实验方案，做实验，将实验数据交给研究员进行分析报告。模板化的实验方案方便实验员理解实验设计。实验室人员不论是否熟悉数据分析，都可以轻松完成实验报告。大大提高实验成功率和跨部门合作效率。  
Event has ended
JMP在半导体制造工艺差异性及过程能力分析中的应用 – 应用材料
Thursday, May 30, 2024
NA
在半导体制造领域，各种工艺（如蚀刻）需要被集成在一起。在量产过程中，多台相同类型的机台需要用在同 一道工序。因此，消除机台间性能差异、提高工艺能力和稳定性至关重要。本项目的主要目标是为了验证新机 台（腔室）是否与标准机台（腔室）匹配，并评估当前工艺的能力和稳定性。 在机台匹配部分，我们首先确定合适的子组，并验证它们是否属于正态分布。之后，对于正态分布群或样本均 值和标准差可以信任的非正态群，采用单因素方差分析进行组均值方差检查。接着用JMP工具（如多重比较和 等效测试）找出哪个机台与标准机台有显著差异，以及它们之间的差异有多大。 此外，回归模型还可用于分析响应与潜在因素之间的相关性。通过线性或多项式拟合，我们可以找出机台之间 存在差异的原因。另外，我们还可以使用配对 t 检验（不需要正态性）来分析配对差异并将结果与方差分析模 型进行比较。 在过程能力和稳定性分析中，我们首先需要找到不属于正态分布的子群的样本均值和标准差的替代值。例如， 我们通常可以使用鲁棒均值和鲁棒标准差来替代单侧或双侧异常值问题。最后，我们通过Ppk、Cpk和稳定性 指数来分析了工艺能力的稳定性。我们还用了目标图和过程稳定性性能图来帮助结果更好可视化。
Event has ended
利用JMP来评估二氧化硅膜厚的工程能力指数和稳定指数以及在膜厚变化中的归因分析 – 应用材料
Thursday, May 30, 2024
NA
在半导体芯片生产中，快速热退火的腔体内壁会由于长时间加工产品而受到覆盖物（Coating）的影响，从而影响内部 控温，因此需要定期通过快速热处理后晶圆的二氧化硅厚度或者阻值来检测腔体的健康程度，决定是否进行预防性维修 （PM）。我们发现现有机台二氧化硅厚度的稳定性与Golden tool存在差距，但却无法被直接量化，且归因分析较为复 杂，于是利用JMP量化腔体的工艺能力并辅助归因分析，最终获得了比Golden Tool更好的表现。我们首先利用IMR控制 图、过程表现图以及目标图可视化了腔体的工艺能力，随后通过计算过程能力指数以及稳定指数清晰地量化了腔体的工 艺能力。随后，利用多因子模型分析探究了现有机台与Golden tool之间WtW以及WiW均一性的差别。利用XBar-S控制 图、过程能力图对比了现有机台与Golden tool二氧化硅厚度和退火后阻值的表现，显示退火后阻值均一性远优于 Golden tool，更稳定更有能力，因此可以排除机台硬件的因素。再经过进一步的分析后，发现此种膜厚的变化在 Golden tool也潜在存在。在经过工艺调试后，二氧化硅厚度的稳定性得到了极大的改善，并且调试过后的阻值均一性、 稳定性与工艺调试前匹配一致。过程表现图、目标图以及过程能力指数、稳定指数的对比，也同样显示出此种工艺调试 的效果切实可行。
Event has ended
表面粗糙度的数据挖掘分析 – 应用材料
Thursday, May 30, 2024
NA
表面粗糙度是半导体和显示器行业中衡量缺陷程度的关键指标，这是因为表面粗糙度会降低器件的电学性能且 影响器件使用寿命。在加工过程中，造成表面粗糙度工艺失效的原因可能各不相同，但由于缺乏能全面评估与 分析表面粗糙度指标的有效模型，使我们在改进工艺方面面临巨大挑战。 本项目旨在建立 JMP 分析包，将表面粗糙度测量指标与非正态分布模型建立联系，以便进行高效的根本原因 分析和工艺调整。 本项目使用 JMP Random Simulation平台生成粗糙度 Z-profile 数据并分为六种非正态分布模型：(1) normal; (2) uniform; (3) heavy tail; (4) right skewed; (5) bimodal; (6) outliers (3%). 计算了五 个表面粗糙度测量指标(Ra, Rz, Rp, Rv, Rk)和八个描述分布性的统计参数。比较了不同的clustering聚类方 法，包括Hierarchical Cluster, K-means Cluster, Normal Mixtures, Cluster Variables, Multivariate Correlation, Principle Components Analysis和 Model Driven Control Charts，看它们是否能有效地将 六个非正态分布模型分为轻尾聚类和离散点，并将 13 个变量分为峰值聚类、非对称聚类和轻尾聚类。有效区 分并预测表面粗糙度测量指标与非正态分布模型之间的关系。 JMP 表面粗糙度Data Mining项目能有效地检测工艺失效原因，进而缩短了工艺调试时间。此外，我们还在公 司内部搭建了一个强大的跨部门 JMP 表面粗糙度分析团队，为整个应用材料公司建立数据库（粗糙度、原始 Z-轮廓、粗糙度衡量、工艺调整）和开发预测模型。
Event has ended
理解和探索表面粗糙度的特性用于增强归因分析与工艺调节在显示面板行业中的应用 – 应用材料
Thursday, May 30, 2024
NA
表面粗糙度是半导体和显示器行业中衡量缺陷程度的关键指标，这是因为表面粗糙度会降低器件的电学性能且 影响器件使用寿命。在加工过程中，造成表面粗糙度工艺失效的原因可能各不相同，但由于缺乏能全面评估与 分析表面粗糙度指标的有效模型，使我们在改进工艺方面面临巨大挑战。 本项目旨在建立 JMP 分析包，将表面粗糙度测量指标与非正态分布模型建立联系，以便进行高效的根本原因 分析和工艺调整。 本项目使用 JMP Random Simulation平台生成粗糙度 Z-profile 数据并分为六种非正态分布模型：(1) normal; (2) uniform; (3) heavy tail; (4) right skewed; (5) bimodal; (6) outliers (3%). 计算了五 个表面粗糙度测量指标(Ra, Rz, Rp, Rv, Rk)和八个描述分布性的统计参数。比较了不同的clustering聚类方 法，包括Hierarchical Cluster, K-means Cluster, Normal Mixtures, Cluster Variables, Multivariate Correlation, Principle Components Analysis和 Model Driven Control Charts，看它们是否能有效地将 六个非正态分布模型分为轻尾聚类和离散点，并将 13 个变量分为峰值聚类、非对称聚类和轻尾聚类。有效区 分并预测表面粗糙度测量指标与非正态分布模型之间的关系。 JMP 表面粗糙度Data Mining项目能有效地检测工艺失效原因，进而缩短了工艺调试时间。此外，我们还在公 司内部搭建了一个强大的跨部门 JMP 表面粗糙度分析团队，为整个应用材料公司建立数据库（粗糙度、原始 Z-轮廓、粗糙度衡量、工艺调整）和开发预测模型。
Event has ended
半导体元件性能预测：模型优化探索 – 应用材料
Thursday, May 30, 2024
NA
半导体制造过程中，预测元件性能面临着挑战，这种不可预测性会导致产量下降和成本增加。本报告旨在建立一个预测模型，利用线上检测数据（包括3个响应变量（Y）和300多个预测变量（X），数据集约500个数据点）来预测元件的性能。   使用JMP多变量平台对项目进行评估，结果显示项目具有相对较弱的Y-Y竞争、强烈的Y-X因果关系，以及严重的X-X多重共线性。考虑到所有300多个预测变量时，标准最小二乘和逐步算法均显示出严重的过度拟合风险。因此，我们运用了四种数据挖掘技术，从中选择了少数关键的X变量进行RSM建模，包括：响应筛选、分区、神经网络和预测变量筛选。由于仅使用常规JMP软件，我们以预测变量筛选的贡献模式和排名作为评估其他三个数据挖掘平台的参考。其中，预测变量筛选算法（随机森林）能解决递归分割问题，使贡献模式更加均匀。   基于预测变量筛选的贡献模式，我们运行了适应模型、神经网络和分区模型来预测三个设备的性能。这个JMP项目成功地提升了我们的预测能力，并从增强现实应用中获取了有价值的见解。同时，可靠的预测模型也能显著缩短新产品开发周期。本研究的结果可为半导体制造业提供重要的参考，以改善元件性能的预测能力。
Event has ended
JMP GRR 项目: GRR在半导体工艺和量测机台的应用研究 – 应用材料
Thursday, May 30, 2024
NA
量具重复性和再现性(GRR)广泛用于评估测量仪器的精度。然而，GRR分析在工程工作中往往被忽视。因此， 为了提高工程质量和效率，必须实施系统的GRR研究。本文重点介绍了GRR在应用材料量测机台和工艺机优化 中的创新应用，并成功解决了客户的HVP。 1: EPI电阻率测量工具的GRR分析：电阻率（RS）是EPI薄膜的关键参数，当前测量系统的稳定性严重影响 EPI工艺的鉴定和开发。经过GRR分析，P/T ratio为 172%（>30%）。后续分析表明，GRR 性能不佳的原 因是工具测量重复性差。接下来对不同时间测试的parts进行Pair-T 测试并创造性使用IMR图对晶圆内外圈的 电阻进行分组分析后，发现重复性差的原因是晶圆在测试过程退化以及晶圆内外氧化层厚度和电荷均匀性不 同，而这总差异是由当前RS测量系统造成的。基于以上实验分析，优化测量工具的改进方案提出，后续的优化 正在进行中。 2: GRR在晶片光学检测设备捕获率（CR）监控方面的应用：目前晶片光学检测设备捕获率监控方法没有被 qualify。因此，采用 GRR 分析来qualify当前监测方法有效性和工具在月保期间的漂移十分必要。结果表 明：当前的监控方法是有效的（P/T=17%），机台在月保期间没有漂移问题。项目同时对P/T ratio较大（> 10%）的根本原因进行了分析。结果表明：由于机台自身的能力限制，机台对特定的100nm 突起和侵入缺陷 检测上存在检测重复性问题。通过这项研究发现：仅对缺陷数量进行GRR分析不足以监测当前机台性能，需要 对最具挑战性的缺陷类型进行Attribute GRR分析，这将是学检测设备性能监测的新方向。 3：测量时间对二氧化硅厚度影响的GRR分析：热氧化产生的二氧化硅厚度会随着时间的推移而增加，如果测 量时间选择不当，会严重影响测量的结果。经过GRR 分析发现，目前的测量时间间隔和方式满足要求（P/T ratio=9%）。然后进一步对提高工艺规格的可能行进行了研究,结果表明：在当前的测量GRR能力下，工艺规 格可以进一步提高，缩小到原来的40%，而 此时P/T ratio可以保持在15%。最后，通过 ICC、P/T 和 Cp 图表平台，开发了一种持续改进流程和GRR能力的系统方法，为持续优化GRR和工艺提供了指导。 GRR项目的成功实施，使整个团队认识到GRR在测量工具监测和工艺优化中的重要性，并且根据以上项目的研 究，团队制定了GRR BKM实施的改进计划，为提高项目的质量和效率奠定了坚实的基础
Event has ended
DOE and Consumer Research: Tackling Consumer Preference Variability
Tuesday, October 22, 2024
Executive Briefing Center 8
One of the great things about humans is that we are all unique. Diversity has benefits all around us, but variation in preferences with regard to consumer goods makes it challenging to predict what will delight the greatest number of people. The use of design of experiments (DOE), facilitated by JMP, is critical to designing formulations with the most appeal.  The objective of this presentation is to interactively present methods to optimize formulations for the greatest consumer liking. Two DOE approaches from a real food formulation case study will be presented: an 18-run definitive screening design (DSD) and an 18-run space-filling design modeled using SVEM and neural networks. To prepare data for modeling, multidimensional scaling is demonstrated to remove anomalous participants’ data. Participant clusters built using hierarchical clustering are used when fitting models using fit definitive screening and SVEM neural networks. The power of the JMP profiler will highlight how consumer preferences differ by cluster. Text Explorer is used to show how to verify insights gained through modeling by exploring verbatim comments by study participants. Lastly, insights gained from each experimental design and modeling approach are compared along with limitations of each. Attendees are presented with a more information-rich alternative to traditional DOE and consumer testing strategies.
Golf is Hard, but Designing Experiments is Easy
Tuesday, October 22, 2024
Executive Briefing Center 8
Once you’ve learned how easy it is to design an experiment in JMP, you never look at the world around you the same. Everything becomes an opportunity for an experiment! This presentation uses a practical example to demonstrate the process of design of experiments (DOE), including designing the experiment, modeling the results, and optimizing the inputs to provide the most desirable output. Attendees at last year’s Discovery conference were treated to an evening of unique fun: hitting glow-in-the-dark golf balls on the driving range at Indian Wells Golf Resort. The driving range has Toptracer technology that monitors each shot. Total distance, carry, ball speed, launch angle, and curve are some of the variables reported with each shot. A driving range that provides so much data provided a perfect opportunity to design an experiment using JMP! After an evening with fellow JMP users and friends, an experiment was designed using the Custom Designer in JMP. The design took only minutes to create. Input variables based on the golf stance setup were used in the design. These included variables such as grip, club head alignment, stance width, and ball location. The designed experiment was executed on the driving range, a model was created, and optimum settings to create the longest and straightest shot were discovered. The modeling and optimization were completed in minutes, while still on the driving range! This allowed for confirmation runs to immediately be performed. The benefits were later transferred the golf course as well.